SoylentNews is people
SoylentNews
from the giving-BSOD-a-whole-new-meaning dept.
Microsoft has vowed to "solve the problem of cancer" within a decade by using ground-breaking computer science to crack the code of diseased cells so they can be reprogrammed back to a healthy state.
[...] The researchers are even working on a computer made from DNA which could live inside cells and look for faults in bodily networks, like cancer. If it spotted cancerous chances it would reboot the system and clear out the diseased cells.
Chris Bishop, laboratory director at Microsoft Research, said: "I think it's a very natural thing for Microsoft to be looking at because we have tremendous expertise in computer science and what is going on in cancer is a computational problem.
[Continues...]
Dr. Lowe, from In the Pipeline, is not convinced that Microsoft is being realistic with their "molecular computer" that will cure cancer:
We're not even near understanding what's going on in normal cells or cancerous ones, so giving people the impression that you've already simulated everything important and you're busy "debugging" it is not only arrogant, it's close to irresponsible.
[...] If you remove the hubris from the Microsoft announcement, though, which takes sandblasters and water cannons, you get to something that could be interesting. It's another machine learning approach to biology, from what I can make out, and I'm not opposed in principle to that sort of thing at all. It has to be approached with caution, though, because any application of machine learning to the biology literature has to take into account that a good percentage of that literature is crap, and that negative results (which have great value for these systems) are grievously underrepresented in it as well.
[...] So if Microsoft wants to apply machine learning to cancer biology, I'm all for it. But they should just go and try it and report back when something interesting comes out of it, rather than beginning by making a big noise in the newspapers. You want to cure cancer? Go do it; don't sit around giving interviews about how you're going to cure cancer real soon now.
Note: Bold added by submitter.
http://www.telegraph.co.uk/science/2016/09/20/microsoft-will-solve-cancer-within-10-years-by-reprogramming-dis/
http://blogs.sciencemag.org/pipeline/archives/2016/09/21/better-faster-more-comprehensive-manure-distribution
(Score: 5, Informative) by Anonymous Coward on Thursday September 22 2016, @12:09PM
Cancer is not one disease. It's cells which have broken through many different checks and balances in the process of DNA replication and apoptosis/immune destruction of things that go out of control. These checks and balances are pretty good, which is why the trillions of cells that make up you are doing OK. But 37 trillion rolls of the dice multiplied by how many you shed and replace often is a hell of a lot of rolls of the dice for something to go wrong.
When you think about it this way, the real surprise with cancer is that it's not more common.
Here's the rub. When this process goes off the rails badly enough to create what we call cancer, most of those checks and balances are completely shredded. That means it goes further and further off the rails. Even in a single person, "cancer" is not a single disease. It's not one mutation. The tumor has a crazy number of different mutations, often even a wildly divergent number of chromosomes within the same mass.
There are rare exceptions to this, like the canine cancer spread by sexual contact in dogs. That one for some reason is stable. So there are exceptions that prove the rule. However, in the general case there is no particular mutation and the entire approach posited is fallacious from the start, because it isn't one disease.
If that wasn't enough: "Living computer made from DNA" just doesn't pass the sniff test, either, and rebooting cells is a much worse idea than just killing wayward ones. Good luck with either. Cancer is great at hiding from the immune system, so their supposed solution would have to be separate and superior to our immune system... which we do not even fully understand. That is not going to happen.
(Score: 3, Interesting) by opinionated_science on Thursday September 22 2016, @12:34PM
I agree. Not anon. Just wrote up a summary of 25 genetic diseases for $PROJ, and every single one is f*ing horrible.
When the machine breaks, all you can hope is that you don't feel the pain.
Probably the more relevant question for "cancer", is the role of the various actors - DNA, repair mechanisms , control machinery (methylation, PTM's, protein efficiency) , environmental exposure, infection (virus/microbe/other)...
And the fact another (google also declared "cure the world" in a press release) computing company is saying this, suggests there is money to be siphoned from gullible investors - our collective pension funds, maybe?
(Score: 1, Interesting) by Anonymous Coward on Thursday September 22 2016, @02:07PM
The same results can only be arrived by others a very low percentage of the time (~10%) when it comes to cancer research, and it is to the point where peopole are actively resisting attempts by others to replicate their experiments. Further, people trying to do these replications are quitting in disgust saying these studies are so crappy it isn't even worth repeating them. Even the people who don't like trying to reproduce cancer research say it is because it is naive to clean out the "Augean Stables" that way.
http://www.slate.com/articles/health_and_science/future_tense/2016/04/biomedicine_facing_a_worse_replication_crisis_than_the_one_plaguing_psychology.html [slate.com] [slate.com]
http://www.nature.com/nrd/journal/v10/n9/full/nrd3439-c1.html [nature.com] [nature.com]
http://www.reuters.com/article/us-science-cancer-idUSBRE82R12P20120328 [reuters.com] [reuters.com]
http://www.nature.com/nature/journal/v483/n7391/full/483531a.html [nature.com] [nature.com]
http://www.sciencemag.org/content/348/6242/1411 [sciencemag.org] [sciencemag.org]
http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002165 [plos.org] [plos.org]
http://www.nature.com/news/cancer-reproducibility-project-scales-back-ambitions-1.18938 [nature.com] [nature.com]
I wouldn't take that literature seriously if I were you, it is probably worse than psychology. I haven't looked too much into it yet, but suspect virology will win out as the absolute bottom of the barrel of research fields in the end, they fail to blind themselves in nearly every paper I have read.
(Score: 1, Interesting) by Anonymous Coward on Thursday September 22 2016, @01:58PM
This cancer is many diseases thing is a cop out meant to make you excuse the total lack of progress that has been made on the topic. Cancer is one disease, if you could detect and selectively remove aneuploid cells from tissues where they do not belong, you could cure 90-99% of cancers.
Anyway, this story is just about detecting cancer, not even killing/removing/controlling it. As mentioned above, we already know how to tell if a cell is cancerous, you check the number of chromosomes (this is called karyotyping). This has been known for 100 years or so. It is pretty idiotic to say the problem will be "solved" even if it worked.
(Score: 2, Interesting) by Anonymous Coward on Thursday September 22 2016, @04:23PM
I am the GP.
Not only is cancer many diseases, but - with the most select of exceptions - cancer is unique in every individual with the disease. Despite this, we've actually made significant progress in treating many cancers. However, therapy != cure and as I articulated above, thanks to rolling the dice every living thing will remain susceptible to cancer if they live long enough.
Karyotyping detects some cancers. Not all. However, even if I accepted that point, the immune system has to work with what it can 'see', which means proteins/lipid bilayers on the surface of cells, or molecules/proteins those cells are excreting/secreting. It's easy to sit in your armchair and say "hey, just remove cells with the wrong number or arrangement of chromosomes" but there is literally no way to know this as said chromosomes are secreted away in the nucleus.
Seems easy, until it isn't.
(Score: 0) by Anonymous Coward on Thursday September 22 2016, @04:52PM
No, it really is one disease defined by the presence of aneuploidy. There have been tens of thousands of papers published on this, it has even trickled down to wikipedia:
https://en.wikipedia.org/wiki/Aneuploidy [wikipedia.org]
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3675379/ [nih.gov]
The earliest article I saw on pubmed:
http://www.nature.com/articles/184290a0 [nature.com]
And you are correct on this point:
As I said, we have known how to identify cancer cells for a long, long time (via karyotyping). That is not the problem. The problem is identifying the aneuploid cells in living tissue and selectively removing them. That is where the funding should be going, not into all this "cancer is many diseases" junk. It is one disease with one commonality. That is really one of the most damaging ideas I have ever heard, it will do nothing but lead to misdirected resources.
(Score: 3, Informative) by Joe on Thursday September 22 2016, @05:47PM
That is where the funding should be going
Sounds like you've got a perfect solution (Nirvana fallacy).
Despite what you say, there has been a lot of real progress made in the treatment of cancer. Some of the therapies are specific to certain types of cancer (Imatinib, Tamoxifen, and Rituximab come to mind) and others such as immune checkpoint inhibitors are showing a lot of progress across multiple types of cancer.
- Joe
https://en.wikipedia.org/wiki/Nirvana_fallacy [wikipedia.org]
(Score: 0) by Anonymous Coward on Thursday September 22 2016, @07:04PM
I had an earlier response that didn't show up. In short:
1) You are attacking a strawman, I never said it would work for sure. I said funding should be focused on that, which is no fallacy. It is totally rational to focus your efforts on targeting one universal aspect of cancer that has been consistently reported for nearly a century using many different technologies. Apparently the alternative is to make millionth-assed attempts at a million different flash-in-the-pan targets, since the disease does not manifest in exactly the same way every time. That sounds like a great way to waste as much time and money as possible before figuring anything out.
2) If you can link to the evidence that has convinced you of the real progress I will look into it. From other posts in this thread you can see the standards used in cancer research are extremely low, so we need to check it for ourselves.
(Score: 2) by Joe on Thursday September 22 2016, @04:37PM
This cancer is many diseases thing is a cop out meant to make you excuse the total lack of progress
No, people say that "cancer isn't one disease" to highlight the vast differences among types of cancer. Different types of cancer can cause different pathology and require different types of treatment (anti-angiogenesis drugs and surgery would be useless for lymphomas).
Cancer has many commonalities and aneuploidy is one of them, but cancer does not require aneuploidy (inactivation of tumor suppressors and the presence of oncogenes can drive cancer independent of chromosomal number).
I consider Cancer to be a class of diseases - similar to Autoimmunity (Lupus, arthritis, multiple sclerosis, psoriasis, Crohn's, and APECED are all very different).
- Joe
https://en.wikipedia.org/wiki/The_Hallmarks_of_Cancer [wikipedia.org]
(Score: 2) by Non Sequor on Thursday September 22 2016, @04:39PM
From a physics perspective, you can treat the problem of identifying cancerous cells as reversing an increase in entropy due to mixing bad cells with good ones.
If there are easy to exploit chemical (differential in response to a toxin) or mechanical (confined to a discrete mass of tissue) distinctions between good and bad cells, then you have a cancer treatment. Separating a mixture based on its properties is analogous to using water to separate salt and sand or using a centrifuge to separate a mixture by density.
But if you lack strong differentiators, you're left with a mixing entropy which requires a quantity of physical and computational work to eliminate.
Write your congressman. Tell him he sucks.
(Score: 0) by Anonymous Coward on Saturday September 24 2016, @10:02PM
This is a very apt description of the problem. Thank you.